Scholarlink

<p>Fused deposition modeling (FDM) printing is widely used for polymer manufacturing. However, the mechanical properties of the final parts, such as tensile strength (TS), are critically dependent on printing parameters. Manually determining the optimal parameter setting is challenging. This study seeks to analyze the effects of the printing parameters on the TS of FDM-printed polyethylene terephthalate glycol (PETG) parts and optimize the printing parameters using statistical analysis and machine learning (ML) approaches. A fractional factorial experiment was designed using Taguchi L27 orthogonal array, considering five critical process parameters. The effects of these parameters on the TS of PETG specimens were investigated, and 27 predictive models were developed and benchmarked. Through experimental data analysis using signal-to-noise (S/N) ratio and analysis of variance (ANOVA), it was identified that layer thickness, build orientation, and printing speed as the statistically most influential factors affecting TS. The gradient boosting regressor model achieved the lowest prediction errors and the highest coefficient of determination (R<sup>2</sup> = 0.923). Furthermore, explainable artificial intelligence (XAI) was employed based on shapley additive explanations (SHAP) to interpret feature influence, ensure model transparency, and optimize the process parameters. This approach provides a robust framework for manufacturers to predict and optimize the TS of FDM-printed PETG components, enhancing part quality and manufacturing efficiency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Analysis and optimization of FDM printing process of PETG parts using statistical and machine learning methods

  • 期刊论文

  • 发表日期：

    2026年1月19日

查看全文
  • Anamul Hasan,
  • Synthia Ferdouse,
  • Mohammad Tafsir Ahmed Shiddiki,
  • Zhong Hu

摘要

Fused deposition modeling (FDM) printing is widely used for polymer manufacturing. However, the mechanical properties of the final parts, such as tensile strength (TS), are critically dependent on printing parameters. Manually determining the optimal parameter setting is challenging. This study seeks to analyze the effects of the printing parameters on the TS of FDM-printed polyethylene terephthalate glycol (PETG) parts and optimize the printing parameters using statistical analysis and machine learning (ML) approaches. A fractional factorial experiment was designed using Taguchi L27 orthogonal array, considering five critical process parameters. The effects of these parameters on the TS of PETG specimens were investigated, and 27 predictive models were developed and benchmarked. Through experimental data analysis using signal-to-noise (S/N) ratio and analysis of variance (ANOVA), it was identified that layer thickness, build orientation, and printing speed as the statistically most influential factors affecting TS. The gradient boosting regressor model achieved the lowest prediction errors and the highest coefficient of determination (R2 = 0.923). Furthermore, explainable artificial intelligence (XAI) was employed based on shapley additive explanations (SHAP) to interpret feature influence, ensure model transparency, and optimize the process parameters. This approach provides a robust framework for manufacturers to predict and optimize the TS of FDM-printed PETG components, enhancing part quality and manufacturing efficiency.